基于log-Gabor小波的掌纹识别算法  

Palmprint recognition based on log-Gabor wavelet

作　　者：赵静岚[1] 庄镇泉[1] 李斌[1] 

机构地区：[1]中国科学技术大学电子科学与技术系,安徽合肥230027

出　　处：《中国科学技术大学学报》2008年第10期1158-1162,共5页JUSTC

摘　　要：掌纹识别是一种新兴的身份识别技术,具有易于采集、纹理丰富等优点,为此提出一种基于log-Gabor小波进行特征提取的掌纹识别算法.该算法首先用log-Gabor小波对掌纹目标区域(region of interest,ROI)进行滤波,然后根据滤波后图像的相位信息形成二进制掌纹特征码,最后用汉明距离来衡量不同掌纹特征码的相似度.在UST掌纹库上的实验结果达到了较高的识别率,验证了该算法的有效性.As a new technology for biometric recognition, palmprint recognition has the advantages of easy image capturing and richness in texture. A novel palmprint recognition algorithm based on log-Gabor wavelet was proposed. First, the ROI central part was filtered by a bank of 2-D log-Gabor wavelets. Then, phase code was generated from the texture information. Finally, two palmprint images were compared in terms of their hamming distance. The experimental result reached a high recognition rate and shows efficiency on UST hand image database.

关 键 词：掌纹识别 LOG-GABOR小波 纹理分析 特征提取 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]